[gmx-users] How to define pull-init for two pull groups

[Justin Lemkul]

On 1/17/14, 8:06 PM, Rini Gupta wrote:
> Dear gmx-users,
>
> Thanks for the reply.
>
> That means if I will use the frames generated by script (setup-umbrella.py)
> it will not generate a sufficient overlap between adjacent
> windows  for this system?

What I'm saying is that I know the scripts, input files, etc. work as advertised 
for the system in the tutorial.  They may or may not work for any other system 
you may create.  That's up to you do evaluate.

> So, In order to extract starting window structures for umbrella sampling of
> my system I prepared a  list of frames at a distance of 0.1 nm
> including frames where distance between reference group and pull group is
> decreasing in steps of 0.1 nm and then increasing by same amount.
>   Please let me know if this is right.
>

Whether or not something is "right" depends on whether or not it works.  In the 
case of umbrella sampling, that means the sampling overlap is adequate.  Again, 
that is for you to judge.

> Then, in order to equilibrate each of these initial configuration, I use
> the following .mdp settings (as given by tutorial)
>
>
>> ; Pull code
>> pull            = umbrella
>> pull_geometry   = position  ;
>> pull_dim        = N N Y
>> pull_start      = yes        ; define initial COM distance > 0
>> pull_ngroups    = 2
>> pull_group0     = cbilayer  ; center of carbon's of bilayer
>> pull_group1     = phosup   ; phosphate atom of 16th PEPC in top
>> pull_pbcatom1   = 0
>> pull_vec1       = 0.0 0.0 -1.0
>> pull_init1      = 0 0 0
>> pull_rate1      = 0      ; no pulling
>> pull_k1         = 1000      ; kJ mol^-1 nm^-2
>> pull_group2     = phoslow   ; phosphate atom of 56th PEPC in bottom
>> pull_pbcatom2   = 0
>> pull_vec2       = 0.0 0.0 1.0
>> pull_init2      = 0 0 0
>> pull_rate2      = 0      ; no pulling
>> pull_k2         = 1000      ; kJ mol^-1 nm^-2
>> pull_nstxout    = 100      ; every 2 ps
>> pull_nstfout    = 100      ; every 2 ps
>>
>>
> After equilibration of 4 ns, if I want to extend these runs  for longer
> time using tpbconv:
>
> # tpbconv_mpi_d -f npt0.trr -s npt0.tpr -e npt0.edr -n index.ndx -extend
> 2000 -o npt0.tpr
> # mdrun_mpi_d -s npt0.tpr -cpi npt0.cpt -deffnm ${MDRUN_RUN_NAME}
>
> ----------------------------------------------------------------------------------------------------------------------
> Program tpbconv_mpi_d, VERSION 4.6.2
> Source code file: /home/roman/gromacs-4.6.2/src/gmxlib/index.c, line: 1192
>
> Fatal error:
> Cannot read from input
> and
> Program mdrun_mpi_d, VERSION 4.6.2
> Source code file: /home/roman/gromacs-4.6.2/src/gmxlib/checkpoint.c, line:
> 2118
>
> Fatal error:
> Can't read 34580 bytes of 'npt0.xvg' to compute checksum. The file has been
> replaced or its contents have been modified. Cannot do appending because of
> this condition.
> ----------------------------------------------------------------------------------------------------------------------------------------------------
> Please let me what is wrong here, I did not modify any file after 4 ns and
> I have been successfully extending  normal simulation runs (without pull
> code) using these settings.
>   So, I guess here problem is related to
> generation of *.xvg file.
> How to supply this file to tpbconv? I have searched on mailing list but
> didn’t get any satisfactory answer for extending equilibration or even
> production runs (which generate pullf.xvg and pullx.xvg files)
> for these umbrella sampling simulations.
>

I seem to recall appending to .xvg files is broken.  Just use the -noappend 
option of mdrun.

-Justin

-- 
==================================================

Justin A. Lemkul, Ph.D.
Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 601
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalemkul at outerbanks.umaryland.edu | (410) 706-7441

==================================================